There are currently no reliable in vitro models for brain tumors that predict drug response in humans. Clinical and research evidence indicates that a small population of cancer stem cells (CSCs) in tumors are primarily responsible for tumor initiation, progression, recurrence, and resistance to therapeutics. These cells have self- renewal capacity and unlimited proliferative potential. Although the targeting of CSCs represents a potential treatment approach, it is very challenging to isolate CSCs from human cell lines or primary cancer specimens since they represent such a small proportion of the entire tumor cell population. Existing methods for isolation and expansion of CSCs are ineffective, cumbersome, expensive, and unreliable. Here we aim to develop clinically relevant and predicative in vitro human tumor models for rapid and low-cost drug assessment for anti- CSC therapy. The proposed research uses an advanced 3D system made of complex scaffold of chitosan and alginate (CA), two naturally occurring polymers that bear proxy structure of glycosaminoglycans, a major component of native extracellular matrix (ECM). These CA scaffolds will serve as a niche to selectively renew and rapidly enrich CSCs. Once the preliminary system is established with optimal structural and mechanical properties that demonstrate CSC renewal, we plan to further improve and fine-tune our tumor model by introducing environmental factors such as biochemical coatings, hypoxia, and human stromal signaling factors into CA scaffolds for human glioblastoma culture. The established optimal CA scaffold model is expected to support the formation of CSC-enriched tumor spheroids from cell lines, primary GBM cells, and freshly resected GBM tissue. A small number of cells from the spheroids implanted in nude mice are expected to form orthotopic tumors and recapitulate GBM phenotypes. GBM is selected as the target tumor for the proposed study due to its fatality and dismal prognosis and our extensive clinical and research experience of GBM treatments. Specific aims of the proposed research are to: 1) determine the role of microstructure and mechanical properties of CA scaffolds on enrichment of CSCs and establish CSC-enriched tumor spheroid models; 2) investigate the role of microenvironmental cues of CA scaffolds in CSC enrichment, and tumorigenic capacities of CSC-enriched tumor spheroids; and 3) use the tumor models in drug tests for personalized cancer treatment and design of therapeutic strategies that specifically target CSCs for GBM therapy. This research will provide a new platform for effectively evaluating potential therapeutic drugs by providing a more accurate and stable tumor microenvironment, thus considerably shortening the time and reducing the cost of drug development. The developed tumor models will also allow researchers and medical practitioners to study molecular mechanisms that regulate self-renewal and differentiation of CSCs, and provide insight into the origin of tumor formation and resistance to treatments.